cards-blt-swin-tiny-patch4-window7-224-finetuned-v2

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.2162
• Accuracy: 0.5022

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.4297	1.0	56	1.1976	0.4933
1.4078	1.99	112	1.1964	0.5011
1.417	2.99	168	1.2025	0.4961
1.4163	4.0	225	1.2295	0.4883
1.4318	5.0	281	1.2330	0.495
1.4383	5.99	337	1.2162	0.5022
1.4212	6.99	393	1.2634	0.4717
1.4346	8.0	450	1.3083	0.4689
1.419	9.0	506	1.2719	0.4806
1.4252	9.99	562	1.3048	0.4911
1.4522	10.99	618	1.2708	0.4794
1.3748	12.0	675	1.3720	0.4383
1.3966	13.0	731	1.3095	0.4594
1.4507	13.99	787	1.2430	0.485
1.4033	14.99	843	1.2728	0.4794
1.3972	16.0	900	1.2611	0.4883
1.4136	17.0	956	1.3166	0.45
1.3992	17.99	1012	1.3103	0.4856
1.3614	18.99	1068	1.3302	0.4422
1.3747	20.0	1125	1.2919	0.4856
1.3868	21.0	1181	1.3166	0.4728
1.3399	21.99	1237	1.3200	0.4672
1.3943	22.99	1293	1.2920	0.4811
1.3635	24.0	1350	1.3109	0.4833
1.3724	25.0	1406	1.3100	0.4644
1.3141	25.99	1462	1.3263	0.4978
1.3576	26.99	1518	1.3307	0.4772
1.3022	28.0	1575	1.3409	0.4978
1.2982	29.0	1631	1.3962	0.4583
1.2657	29.99	1687	1.3329	0.4817
1.3152	30.99	1743	1.2973	0.49
1.2924	32.0	1800	1.3159	0.4833
1.214	33.0	1856	1.3955	0.4833
1.2717	33.99	1912	1.4583	0.46
1.2692	34.99	1968	1.3504	0.4939
1.2127	36.0	2025	1.3784	0.4833
1.1956	37.0	2081	1.4184	0.4817
1.2408	37.99	2137	1.3849	0.4944
1.1699	38.99	2193	1.4298	0.4844
1.1727	40.0	2250	1.4331	0.4772
1.1485	41.0	2306	1.4597	0.4672
1.1668	41.99	2362	1.4429	0.4783
1.1881	42.99	2418	1.4555	0.4839
1.1204	44.0	2475	1.4648	0.4783
1.1523	45.0	2531	1.4744	0.4733
1.1206	45.99	2587	1.4792	0.4906
1.1135	46.99	2643	1.5009	0.4678
1.1227	48.0	2700	1.5480	0.4733
1.1017	49.0	2756	1.5907	0.4644
1.1601	49.99	2812	1.5136	0.47
1.1239	50.99	2868	1.5384	0.4789
1.09	52.0	2925	1.5716	0.4711
1.1023	53.0	2981	1.5736	0.4728
1.1038	53.99	3037	1.5919	0.4556
1.058	54.99	3093	1.5534	0.4772
1.0405	56.0	3150	1.5788	0.4717
1.0172	57.0	3206	1.5855	0.4767
1.0036	57.99	3262	1.6425	0.455
1.0124	58.99	3318	1.6039	0.4678
1.0647	60.0	3375	1.5891	0.4572
1.0143	61.0	3431	1.6265	0.4483
1.0051	61.99	3487	1.6208	0.4633
0.9571	62.99	3543	1.6874	0.4483
0.9838	64.0	3600	1.6778	0.4517
0.9995	65.0	3656	1.6248	0.4722
1.0374	65.99	3712	1.6645	0.4667
0.9483	66.99	3768	1.6307	0.4611
0.9825	68.0	3825	1.6662	0.4661
1.0023	69.0	3881	1.6650	0.46
0.9642	69.99	3937	1.6953	0.4494
0.9687	70.99	3993	1.7076	0.4661
0.9542	72.0	4050	1.7012	0.4656
0.9378	73.0	4106	1.7056	0.4533
0.9542	73.99	4162	1.7331	0.4572
0.9035	74.99	4218	1.7459	0.4417
0.9631	76.0	4275	1.7236	0.465
0.8759	77.0	4331	1.7294	0.455
0.9218	77.99	4387	1.7654	0.4578
0.9077	78.99	4443	1.7234	0.4594
0.8924	80.0	4500	1.7256	0.4683
0.9156	81.0	4556	1.7320	0.4678
0.806	81.99	4612	1.7348	0.4661
0.8863	82.99	4668	1.7514	0.4606
0.8698	84.0	4725	1.7484	0.4661
0.8623	85.0	4781	1.7420	0.4778
0.8643	85.99	4837	1.7636	0.4617
0.8914	86.99	4893	1.7552	0.465
0.837	88.0	4950	1.7552	0.4644
0.8217	89.0	5006	1.7532	0.4639
0.8601	89.99	5062	1.7447	0.4683
0.8293	90.99	5118	1.7622	0.4611
0.8301	92.0	5175	1.7616	0.4633
0.7752	93.0	5231	1.7585	0.4722
0.8533	93.99	5287	1.7842	0.4617
0.8156	94.99	5343	1.7837	0.4622
0.8094	96.0	5400	1.7896	0.4583
0.839	97.0	5456	1.7835	0.465
0.839	97.99	5512	1.7883	0.46
0.7763	98.99	5568	1.7838	0.4594
0.8186	99.56	5600	1.7837	0.4606

Framework versions

• Transformers 4.37.2
• Pytorch 2.0.1+cu117
• Datasets 2.17.0
• Tokenizers 0.15.2